4 ms·
Two things that set it apart: 1. R1 chip, which means that the spatial tracking basically never fails and is way more stable. Typical VR headsets will start la
by phailhaus 3y ago
Two things that set it apart:
1. R1 chip, which means that the spatial tracking basically never fails and is way more stable. Typical VR headsets will start lagging with the more CPU load they have, which causes disorientation and dizziness.
2. Eye tracking and gesture controls. There are cameras point at your eyes, down, and far to the sides. This means that no matter how you're sitting or where your arms are, it can see you "click" with your fingers. This level of seamlessness is necessary for the headset to feel effectively invisible. Typical VR headsets require you to deal with external controllers, which makes them more like consoles. Makes it more awkward to do simple things like check your phone or grab a drink.
I honestly don't think it's going to take off because nobody has demonstrated a "killer app", but the AVP is substantially different from other headsets in the space.
- jayd16 3y agoRE: 1. Do the pico headsets do this or something? Who doesn't have their spatial tracking process set as the highest priority? Are you confusing dropped frames with spatial tracking loss? Apple hasn't invented a technology to never drop frames so I expect once we hop into fully immersive Unity apps we'll have familiar stuttery moments.
- 1123581321 3y agoThere are two processors. The R1 is a realtime processor that manages the vision tracking and passthrough rendering. The M2 is an ARM processor that handles general computing. Processing a Unity app would happen on the M2, and is where frames would be dropped if they aren't ready in time. The R1 should not miss a frame of input or drop a frame from the outside world unless caused by a hardware flaw.
- jayd16 3y agoMy point is that you can still get nauseous from chuggy app rendering even if the passthrough is flawless. Any change in perspective will require an app draw and even if Apple is able to reproject old frames of your app, they will be warped and disorienting. They can't really do enough inpainting to make it flawless.
- 1123581321 3y agoI'm not sure. Inpainting doesn't seem like right way to think about the R1. My understanding is that visionOS uses a rendered mode that gives the application limited control once they've sent the data. That lets the system re-render the application's stale data correctly. In exchange, apps have limited control over the render (whatever can be translated into a subset of MaterialX.) That said, I know Unity's Polyspatial does have a flag to update the draw for each eye individually, so it seems like some disorientation should be possible, but it would be coming from inconsistent data across the eyes, not the way the system processes your head's orientation or the passthrough.
- jayd16 3y agoFull screen Unity apps don't use the MaterialX system. Widgets perhaps work like how you say but VR games and apps won't.
- 1123581321 3y agoWhere do you see that? I thought Unity was having developers handle all of it through the URP via Polyspatial which does convert it. I can see how you could write a bad full-screen app though. Many ways to disorient in addition to inefficient draws. The app store approval process is only going to help so much there.
- jayd16 3y agoURP and their shader graph is hardly a recipe for performance. There would have to be much more than that going on to protect devs from attempting shaders that are too expensive. Full apps don't appear to use the Polyspatial system but its hard to tell exactly.